See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/238186267

Laboratory work in a vehicle dynamics course

Article in Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering · May 2004
DOI: 10.1243/095440704774061156

CITATION READS

1 517

2 authors:

Johan Wideberg Publio Pintado

University of Seville University of Castilla-La Mancha
32 PUBLICATIONS 737 CITATIONS 66 PUBLICATIONS 1,255 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Johan Wideberg on 12 December 2013.

The user has requested enhancement of the downloaded file.

 1

Laboratory work in a vehicle dynamics course

J P Wideberg1* and P Pintado2

1Transportation Engineering, School of Engineering, University of Seville, Sevilla, Spain
2Mechanical Engineering, School of Engineering, University of Castilla-La Mancha, Ciudad Real, Spain

Abstract: This paper describes two experimental facilities and a piece of software designed to serve
as a teaching aid for automotive engineering. The first experimental rig is an antilock brake system
(ABS) workbench for a scaled-down heavy-vehicle tri-axle configuration. The second facility is a
tyre-testing sliding table used to show the forces and moments on the tyre–road contact area. In the
third case, a piece of software has been embedded in an easy-to-use graphic user interface (GUI ) to
analyse vehicle cornering behaviour as predicted by the popular ‘one-track model’. In the first two
cases the focus lies on the construction, on the instrumentation and on the use of the experimental
set-up, whereas in the third case the focus is centred on the equations used to represent the model,
on the simplifications made, on the properties used, on the numerical integration and on the layout
of the designed GUI.

Keywords: ABS, tyre behaviour, one-track model

NOTATION system itself. The students would benefit from experi-

menting with different vehicle velocities, different input
a lateral acceleration (m/s2) braking pressures and different vertical loads. The effects
y
C cornering stiffness, front tyre (kN/rad) of these parameters on the dynamics of braking, with
af
C cornering stiffness, rear tyre (kN/rad) and without an ABS in the control loop, can be easily
ar
DOF degree of freedom assessed and understood using the ABS workbench that
F centrifugal force (N ) is proposed and described in this article.
c
F reaction force at the front tyre (N ) Tyre behaviour is probably the single most influential
yf
F reaction force at the rear tyre (N ) factor in vehicle dynamics. To reinforce the meaning
yr
I moment of inertia (mm4) of this statement the students need to see the different
z
K understeer gradient modes of tyre deformation for themselves and also
v critical speed (m/s) need to be able to ‘feel’ the relationship between tyre
crit deformation and contact forces. The authors believe that
a slip angle of the front tyre (rad) the sliding table for tyre testing described in section 3
f of this article is an excellent tool for teaching tyre
a slip angle of the rear tyre (rad)
r behaviour. The table is a scaled-down version of common
tyre testing rigs available in a number of automotive
1 INTRODUCTION engineering laboratories, with a specially designed sup-
porting arm that plays the role of a home-made load cell.
No secret would be revealed if it were said that Vehicle cornering behaviour may be described by
laboratory work is essential in the learning process simplified models like the ‘one-track model’, also known
for technological subjects, in general, and automotive as the ‘bicycle model’. Despite the simplicity of the
engineering, in particular. In order to understand fully model, it is usually time consuming to integrate the
the working principles behind an antilock brake system dynamic governing equations for certain steering inputs,
(ABS), for example, it is useful to experiment with the whereas, on the contrary, numerical integration is easily
performed using any of the powerful computer tools
available nowadays. Furthermore, a graphic user inter-
The MS was received on 6 November 2002 and was accepted after face (GUI ) may be designed so that the student can
revision for publication on 21 January 2004. easily carry out numerical experiments. The authors have
* Corresponding author: Transportation Engineering, School of
Engineering, University of Seville, Camino de los Descubrimientos s/n, implemented one such ‘numerical workbench’, which is
41092 Sevilla, Spain. described and discussed in section 4 of this paper.
D13802 © IMechE 2004 Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
2 J P WIDEBERG AND P PINTADO

The three ‘experimental’ facilities mentioned in the The fact that an automobile is a moving machine com-
previous paragraphs are used in automotive technology plicates the set-up of experiments related to the vehicle
courses in either the University of Seville or the University itself or to any of its subsystems. Testing the braking
of Castilla–La Mancha. In the case of the University of system of a vehicle, for example, may require a dedicated
Seville (where both authors worked together until just a test track as well as sophisticated on-board equipment.
few years ago), the facilities are used as part of a set of On the other hand, if it were possible to keep the vehicle
interconnected experiments used to train the students in stationary, the experimental set-up would be consider-
understanding vehicle issues related to vehicle dynamics, ably simplified. Eliminating the forward velocity of the
power plant technology, tyre technology and suspension vehicle while simulating the moving road underneath is
and steering system integration. This laboratory work is done in many instances in automotive testing. In the case
part of an automotive technology course that has been of brake testing, however, the added complication of
taught since 1985. It is believed that the course provides simulating the vehicle longitudinal inertia needs to be
the European automotive industry with highly trained solved. In the present rig, this is done by transforming
engineers who are well suited to work in any of the the road into rotating drums whose added moment of
branches of this industrial sector. inertia with respect to their respective axis of rotation
(its respective axles) equals the mass of the vehicle times
the drum’s radius squared. This condition also implies
2 ABS WORKBENCH that the added kinetic energy of the rotating drums is
that of the vehicle.
2.1 Purpose of the experimental rig The experimental workbench has been constructed
The purpose of the experimental rig is to be able to show in collaboration with a local component manufacturer
how the braking system of a heavy vehicle works; to be whose main business lies on the field of semi-trailer
able to show the effects of excessive braking pressure pneumatic brakes. This led the authors to design a test
without an antilock brake system (ABS) in the con- rig suitable for typical three-axle configurations. How-
trol loop; to be able to measure vehicle speed, wheel ever, this decision made it desirable to scale down the
slip, pseudo-slip and pressure near the air actuator size of the wheels and drums since a huge device would
chambers, in order to be able to show their variations be uncomfortable to operate. An effort has been made
and fluctuations during the braking process; and mainly to scale down all relevant parameters in accordance with
to be able to show the effects of an ABS control. These dimensional analysis. Nevertheless, it is clear that most
purposes are achieved with the workbench depicted in of the pneumatic components must be used in their
Fig. 1, where a three-axle configuration is mounted on original size. Moreover, it was not easy to maintain
powered rotating drums that play the role of the road. simultaneous similitude in the case of vertical load and
The drums are brought up to ‘cruising’ speed by an vehicle mass. Nevertheless, since the purpose is to show
electric motor and are left to rotate freely until the the performance of the system, rather than to pre-
braking system is activated. A dedicated data acquisition dict specific vehicle behaviour, the lack of complete
system has been assembled to record data from the similitude is not considered to be a major drawback. The
moment the brakes are applied until the vehicle (the same reason may be argued in order to disregard the
drums in this case) stops. effect of load transfer during braking, the influence of

Fig. 1 The ABS workbench

Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering D13802 © IMechE 2004

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
 LABORATORY WORK IN A VEHICLE DYNAMICS COURSE 3

the suspension in the performance of the braking system,

as well as the influence of the tractor vehicle. Load trans-
fer cannot be modelled in a stationary rig like the one
under consideration; to model the tractor vehicle would
be unfeasible and to include suspension effects would
complicate the deign considerably, but these effects may
be disregarded when the purpose is to show the working
principles behind a braking system.

2.2 General description and construction details

The finished workbench is shown in Fig. 1. The six
drums that simulate the road have been constructed
using truck wheel rims. Each pair of rims is fixed to Fig. 3 Wheel mounting
a rigid axle which, in turn, is joined to the frame via
ball bearings. All three drum axles are connected to an
electric motor with transmission belts and pulleys as
shown in Fig. 2. This motor brings the drums and wheels WABCO Vario C ABS. The system has six sensors (one
up to speed before a test is performed. The electric power for each wheel ) and three modulator valves: MA for the
supply is disconnected just before the brakes are applied. rearmost axle, MB the two front wheels on the right and
A frequency changer allows for test speed selection, MC for the two front wheels on the left. These valves
while a toothed ring (which can be seen in Fig. 2) allows modulate the pressure that governs the air actuators
for accurate measurement of motor speed which, as has by responding to the electronic control unit (ECU )
been mentioned, is proportional to vehicle speed. commands. Needless to say, the system operates as a
It can be seen in Fig. 3 that each wheel is attached to conventional one when the ECU is turned off. The relay
its toothed ring and to its disc brake, and the assemblage valve, the air reservoir and the brake calipers are shown
is mounted on a rigid axle using roller bearings. in Fig. 6.
Therefore, all six wheels could have different angular
velocities in a given dynamic situation. The vertical
position of the rigid axle (to which the wheels are joined) 2.3 Sample experiments
is controlled by means of a threaded strut on either side
Two sample experiments are carried out to show the
of the axle (Fig. 4). These screws are used to press the
ABS workbench at work. The first experiment corre-
tyre against the drum. By adjusting the torque of the
sponds to a sudden application of the brakes with the
corresponding nut, the vertical load on the tyre can be
control unit power supply turned off. Figure 7 shows the
controlled.
vehicle speed versus time which, in this case, is measured
The pneumatic configuration as well as the schematics
as the drum’s angular velocity times its radius. The
of the control system are shown in Fig. 5. The brake
vehicle takes about 12 seconds to stop. It will be seen
system corresponds to that of a semi-trailer with
in the next experiment that this stopping time can be
improved with the ABS turned on. Figure 7 also shows
the angular velocity of one of the six wheels. This
velocity is multiplied by the wheel’s radius, as is usually
done, to be able to plot it on the same scale as the vehicle
velocity. It is seen in Fig. 7 that the wheel rotation stops
soon after the brakes are applied. The vehicle moves for
more than 10 seconds with the wheels blocked, which is
tantamount to saying that the wheels have no directional
preference and that the dynamic behaviour of the vehicle
could be dangerous. Figure 8 clearly shows that the con-
trol pressure increases rapidly after the brake is applied
and that the pressure that reaches the air actuator closely
follows this command.
A very similar experiment is now performed with the
ABS turned on. Vehicle velocity is plotted in Fig. 9.
The stopping time is now significantly shorter (2.5 s
approximately) than in the previous case. Wheel velocity
Fig. 2 Transmission belts and pulleys tends to depart from vehicle velocity (Fig. 9), but when
D13802 © IMechE 2004 Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
4 J P WIDEBERG AND P PINTADO

Fig. 4 Vertical load control screws

Fig. 5 Semi-trailer brake system configuration

(a) (b) (c)

Fig. 6 (a) Relay valve, (b) reservoir (c) and caliper

slip [defined in the usual manner as S=(v−vR)/V ] saturate. Percentage slip is also shown in Fig. 9. Figure 10
grows significantly (Fig. 9) the control system kicks in, shows how the control pressure is constant whereas the
reducing brake pressure. This makes the wheel speed pressure that reaches the air actuator is modulated by
recover, guaranteeing that the friction force does not the control system.
Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering D13802 © IMechE 2004

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
 LABORATORY WORK IN A VEHICLE DYNAMICS COURSE 5

Fig. 7 Vehicle and wheel velocity without ABS

Fig. 8 Control pressure and brake pressure without ABS

Fig. 9 Vehicle velocity, wheel velocity and slip, with ABS

3 SLIDING TABLE FOR TYRE TESTING forces on a body that give rise to the deformations.
Nevertheless, it is possible to twist the reasoning some-
3.1 Experimental set-up and purpose what. It is possible to fix the deformation kinematics of
the body providing the means for generating whatever
Forces on solid bodies are necessarily related to defor- forces are required to cope with the motion. This is the
mations and it is not possible to have either one without case with the sliding table of Fig. 11. No forces can be
the other. The causality principle implies that it is the applied to a tyre without the tyre deforming, which is
D13802 © IMechE 2004 Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
6 J P WIDEBERG AND P PINTADO

Fig. 10 Control pressure and brake pressure with ABS

to the wheel, the tyre is forced to deform in a defined

manner since the wheel and the supporting arm are
much stiffer.
The positioning angle around the vertical axis deter-
mines the angle between ‘road velocity’ (opposite to
vehicle velocity) and the horizontal projection of the
wheel’s angular velocity vector. This is equivalent to
saying that this positioning angle determines the tyre’s
slip angle. This slip angle translates into lateral forces
that may be measured, as will be explained shortly. The
positioning angle around the wheel’s horizontal longi-
tudinal axis is the camber angle, which also translates
into lateral forces.
The supporting arm is much stiffer than the tyre.
This makes it possible to neglect its deformation when
measuring the already mentioned positioning angles.
However, the small deformation of this stiff arm may
nonetheless be measured by means of strain gauges.
The goal is to determine the resultant contact force
(three components, in general ) and the resultant contact
moment (three components, in general ) generated by the
distribution of pressure and the distribution of friction
on the contact area. This force and this moment can be
easily reduced to a point in the section of the vertical
post to obtain the force and the moment at this section,
the components of which are the two shear forces and the
Fig. 11 The sliding table for tyre testing axial force, in the case of the force, and the two bending
moments and the torsion moment, in the case of the
torque. These forces and moments may be measured
the same as saying that when the kinematics is such that with strain rosettes on the four sides of the vertical post.
the tyre is forced to deform, then forces are ‘generated’
in the tyre and the wheel. The sliding part of the table
3.2 Construction details
in Fig. 11 represents the road underneath the tyre. The
arm to which the wheel is connected allows for arbitrary The table frame is constructed using welded squared
positioning of the wheel with respect to the road. The tubes as shown in Fig. 11. This structure needs to be stiff
angle around the vertical axis and the angle around the enough to ensure that positioning angles (camber angle
horizontal longitudinal axis may be set arbitrarily. When and/or slip angle) do not vary significantly when the
these angles are set and the road moves with respect wheel is under load. Figure 12 shows the linear bearing
Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering D13802 © IMechE 2004

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
 LABORATORY WORK IN A VEHICLE DYNAMICS COURSE 7

Fig. 12 Sliding mechanism

guides that permit the sliding motion of the table top. understanding and retention. Traditionally the student
This motion is powered by a motor underneath the table spends a lot of time on such things as deriving the
top (Fig. 13). The motor pinion is meshed to a rack equations of motion and then programming them using
fixed to the sliding table top. a traditional programming language. This approach
Figure 14 shows the wheel positioning devices. often turns out to be more of a programming problem
Figure 14a shows the device on the vertical post that with compiling, debugging, etc. This takes attention
allows for slip angle selection and Fig. 14b shows the away from the solution and understanding the real
device at the wheel’s axle that permits the camber angle problem, in this case the solution of the equations
to be fixed. The horizontal beam of the supporting arm of motions of the ‘bicycle model’. However, by using
can rotate with respect to the vertical post. This rotation a direct interactive visual approach to learning, it is
can be fixed at any of the predefined closely spaced possible to remove many of the conventional barriers
angles by means of the locking pin shown in Fig. 14a. that hinder effective teaching and learning of seemingly
Likewise, the wheel axle can rotate with respect to the abstract scientific and engineering concepts, by reducing
horizontal post, but this rotation can be fixed by means the barriers and obstacles between the student and the
of the locking pin shown in Fig. 14b. concept.
The model that will be used in the interactive
MATLAB interface is known in the literature as the
4 LINEAR ONE-TRACK MODEL bicycle or one-track model, as described in the intro-
duction. The one-track model assumes a stiff chassis
Experience has shown that active visualization-based with its centre of gravity (CG) lying in the plane of the
learning using interactive programs can greatly enhance road. The angle between the speed vector v and the longi-
tudinal axis of the vehicle is called the slip angle b of
the vehicle, which in general is not identical with the
steering angle d.
The suspension of the vehicle is ignored in this model,
thus neither the roll nor the pitching motion of the
vehicle is evaluated. Therefore the CG is assumed to be
located level with the road. Furthermore. the model is
only accurate for small and moderate lateral accelerations,
since the bicycle model is decreasingly applicable with
growing lateral force. Segel [1] says that it can be used
if the lateral acceleration is not greater than 0.4 g. In
the case of the linearized model it will be valid where
the bicycle model holds and the tyres are more or less
linear—except for the study motion on a road with a
low friction coefficient such as snow or ice on the road,
where the slip angle can also become large for small
cornering forces.
In order to reduce the error caused by neglecting the
pitch of the vehicle, the acceleration in the longitudinal
direction will be minimized in the simulation, i.e. the
longitudinal component of the acceleration ẍ=a =0.
x
From the longitudinal equilibrium it can be deduced
effortlessly that the pitching motion also will become
Fig. 13 Traction motor zero.
D13802 © IMechE 2004 Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
8 J P WIDEBERG AND P PINTADO

(a) (b)

Fig. 14 Mechanisms for wheel positioning

The forces exerted on the two front wheels are 4.2 Dynamic equilibrium
assumed to be indistinguishable; the same holds for the
In this paper the equations are expressed as scalar
rear wheels. This makes it possible to replace the front
equations. It is desirable to express these without intro-
and rear wheels with one wheel in the centre of the
ducing the heading angle so ‘vehicle fix coordinates’ are
vehicle subjected to the sum of the original forces. This
used; otherwise an extra integration would be needed in
wheel will have double the cornering stiffness of the
the solution (to keep track of the heading angle).
original wheel. The model will the simulate a vehicle with
The road is considered to be flat and so the motion
two unsprung wheels in one track (similar to a bicycle,
will be planar. The equilibrium of forces gives (see
consequently the name) with no pitch or roll.
Fig. 15)
There are two coordinate systems used to describe

A B
the vehicle: the global system (X, Y, Z) and the vehicle dV
fixed system (x, y, z) with the x axis pointing in the ma =m x −V V =F +F cos (d)−F sin (d)
x dt y xr xf yr
longitudinal direction of the vehicle and the y axis being

A B
positive to the driver’s right-hand side (according to the dV
ma =m y +V V =F +F sin (d)+F cos (d)
SAE definition [2]). y dt x yr xf yr

(3)
4.1 Solicitation input
and the equilibrium of moments gives
There are two different types of input to the model. The
user can choose either step input or sinusoidal input to dV
I =−F c+F sin (d)b+F cos (d)b (4)
the steer angle (d). The step input is defined as a sudden dt yr xf yf
change of the steer angle at t=0 s; i.e. the steer angle
changes from zero to a value delta given by the user
(or by the default of the interface). The definitions of 4.3 Different tyre models
the step and the sinusoidal inputs are given by
The tyre model plays an important role in the perform-
ance of the vehicle model. With the exception of aero-
G
0, t<0
d= (1) dynamic forces, which are only relevant at higher speeds,
d , t
0 all of the interactions of the vehicle with the external
step
d=d sin (2p ft) (2) environment occur through the tyre. Accurate modelling
amp of the tyre is essential to a precise vehicle simulation.
where d is the amplitude, f is the frequency and t is A common approach to calculate the lateral force
amp
the time. The default is a step change and the sinusoidal generated by the tyre is to use either a linearization
steer can be activated by the user by pushing the radio [as in equation (5) below] or the Magic Formula Tire
button on the interface. Model [3], which gives more precise results for larger
Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering D13802 © IMechE 2004

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
 LABORATORY WORK IN A VEHICLE DYNAMICS COURSE 9

Fig. 15 Diagram showing the equilibrium of forces

slip angles. The formula is based on a function whose Fig. 16 and the mathematical definition is given later
behaviour approximates the shape of the curves obtained in equation (6).
from experimental measurements on tyres. Its parameters
are determined so as to fit the curve to a particular set
of experimental results. A problem in many empirical/ 4.3.1 Linear tyre model
numerical approaches is that it requires a large amount The formulation of a linear tyre model is defined below.
of test data which can be difficult and expensive to obtain The lateral force for the rear and front tyres can be
by testing and is often difficult, or even impossible, to expressed as
find by a literature search. However, due to the nature
of this study, a method will be used which has a simpler F =C a
yf af f
mathematical expression that mimics the most important,
F =C a
overall features of the Magic Formula with fewer para- yr ar r
meters. The method often referred to as the brush model (5)
[4], will be used in this paper. The advantage is that it
is still non-linear but does not require experiments where C and C are the cornering stiffness for the front
af ar
and/or analysis. A graphic presentation can be seen in and rear axles respectively.

Fig. 16 Brush tyre model

D13802 © IMechE 2004 Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
10 J P WIDEBERG AND P PINTADO

4.3.2 Non-linear tyre model X and Y:

The brush model will be used because it is easy to dy
implement. It is defined as =V (13)
dt
F =−C tan (a )l (2−l ) dX
yi ai i i i =V cos y−V sin y (14)
F m dt x y
l= zi , l ∏1, l =1, l >1
i 2C |tan (a )| i i i dY
ai i =V cos y+V sin y (15)
(6) dt y x

where the subscript i stands for either r of f, meaning

rear or front axle respectively. The compatibility will 4.5 Numerical mode: non-linear model
then give the following relationship between angles: The results from the dynamic equilibrium and the
tyre model are combined into a numerical model of
bV+V
tan (d−a )= y behaviour for the vehicle. The governing dynamics
f V
x equations can be expressed as
cV−V
G A B
tan (a )= y dV
y=
1 cV−V
y l (2−l )
r V C atan
ar r r
x dt m V
x
(7)

C A B D
bV+V
+C d-atan y l (2−l )
af V f f
x
4.4 Numerical model: linear model
The numerical model for the linear case will be derived
by inserting the equation for the linear tyre (5) in the
×cos (d)−V V
x H (16)

G A B
equations of equilibrium, (3) and (4). The equations will dV 1 cV−V
= −C atan y l (2−l )c
then become dt I ar V r r
x

A B C A B D
dV bV+V
m x −V V =F +F cos (d)−C a sin (d) +C d-atan y l (2−l )
dt y xr xf af f af V f f
x

H
(8)
×cos (d)b (17)

A B
dV
m y +V V =C a +F sin (d)+C a cos (d)
dt x ar r xf af f where equations (16) and (17) are the governing
equations of the vehicle dynamics with the tyre model
(9) included and the influence of the elasticity. The standard
simplification of small angles is not considered because
dV
I =−C a c+F sin (d)b+C a cos (d)b a closed solution is not necessary and the equations are
dt af r xf af f easy to solve numerically. Equations (13) to (15) are
(10) again needed to find out the global heading angle y and
the global coordinates X and Y.
It is also assumed that the angles are small and that the
longitudinal speed is constant. Inserting equation (7)
and simplifying yields 5 THE INTERFACE

A B
dV C +C C b−C c
m y + af ar V + mV + af ar V=C d The interface is divided into several fields with input to
dt V y x V af
x x the bicycle model, feedback and a result visualization
(11) field (see Fig. 17). The different groups are given as
follows:

A B
dV C b−C c C b2+C c2
I + af ar V + af ar V=C bd A. Vehicle parameter input field. This field has some
dt V y V af default values as an example for the user. These
x x
(12) values can all be changed by editing the values. The
parameters are the location of the centre of gravity,
The following three equations are needed to find out the suspended mass of the vehicle and the cornering
the global heading angle y and the global coordinates stiffnesses of the tyres.
Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering D13802 © IMechE 2004

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
 LABORATORY WORK IN A VEHICLE DYNAMICS COURSE 11

Fig. 17 Layout of interface

B. Input related to the input solicitation are in this field, F. Plot area. In this area the results are visualized. In
i.e. the degree of the step input or the amplitude and the example calculated in Fig. 17 the user selects the
frequency of the sinusoidal solicitation, as well as checkbox in order to visualize the results for the
the forward speed and a field for the end-time of the linear and non-linear model.
simulation. G. Calculation button. This button runs the calculation
C. Field for choosing results to plot. Results can be used and automatically plots the lateral velocity and the
in a variety of ways to analyse vehicle manoeuvring jaw velocity.
motions, handling and stability. A large number of
vehicle input and response variables can be selected
for plotting using the time history and XY formats.
A set of two results will be selected and represented
6 SOME NUMERICAL EXAMPLES
as a plot. In order to plot, the button in field E has
to be pressed.
6.1 Over- or understeer
D. Feedback information field. Here additional infor-
mation about the results is presented and alerts on In the information field the student can see feedback on
the validity of the model are issued. The understeer the most recent simulation. For instance, in Fig. 18 the
coefficient is calculated and displayed; if this is results from three different calculations can be seen. Four
negative then the critical speed is also calculated. different parameters can be monitored in this field. First
The maximum lateral acceleration is calculated and is the understeer coefficient K (see Ia or IIIa in Fig. 18);
us
an alert is given if this acceleration is higher than if the value K is positive it means that the vehicle is
us
0.4g (in that case the results are not valid). For understeered and if the value is negative the vehicle
examples see the next section. is oversteered. If that is the case the vehicle is unstable
E. Plot button. This button plots the result of any result above a certain velocity v , which is automatically
crit
that has been chosen. calculated (see IIIa or IIIb in Fig. 18). The lateral
D13802 © IMechE 2004 Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
12 J P WIDEBERG AND P PINTADO

Fig. 18 Display of different results in the feedback field

acceleration (a ) is calculated in order to find out 6.2 Numerical example of different solicitations
y
whether the simulation is valid or not. For accelerations
higher than 0.4–0.5 the bicycle model will give inaccurate There are two different types of solicitation to the bicycle
results. Therefore the user has feedback on whether or model, step change of the steer angle and a sinusoidal
not it is higher than ay>0.4 (see Ib or IIa in Fig. 18). change. In Fig. 19 the results of these two solicitations
The user can also observe the absolute value of the maxi- were plotted. The vehicle modelled had the same para-
mum acceleration in the field a (see Ic or IIb in meters and the velocity was the same. The step steer angle
y,max
Fig. 18). and the amplitude of the sinusoidal where identical. It

Fig. 19 Response to step versus sinusoidal solicitations

Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering D13802 © IMechE 2004

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077
 LABORATORY WORK IN A VEHICLE DYNAMICS COURSE 13

is interesting to observe that due to the relatively low REFERENCES

frequency (0.1 Hz) of the sinusoidal the minimum value
of the lateral velocity and the maximum of the jaw speed 1 Dixon, J. C. Tires, Suspension and Handling, 1991
are identical to the steady state part of the responses of (Cambridge University Press, Cambridge).
the step steer. 2 Gillespie, T. D. Fundamentals of Vehicle Dynamics, 1992
(Society of Automotive Engineers, Warrendale, Pennsylvania).
3 Pacejka, H. B. and Besselink, I. J. M. Magic Formulate Tire
Model with transient properties. Veh. System Dynamics
Suppl., 1997, 27, 234–249.
4 Riekert, P. and Schunck, T. Zur Fahrmechanik des
ACKNOWLEDGEMENTS gummibereiften Kraftfahrzeugs. Ingenieur Archiv., 1940,
11, 210–224.
5 Segel, L. Research in the fundamentals of automobile
The authors would like to thank Mr Diego de la Cruz
control and stability. In SAE National Summer Meeting,
for his work on the sliding table for tyre testing Atlantic City, New Jersey, 5 June 1956.
and Mr Fernando Iniesta for his work on the ABS 6 Vehicle Dynamics Terminology, SAE J670a, 1965 (Society
workbench. The financial support of Jalair S.A. is also of Automotive Engineers, Warrendale, Pennsylvania). An
acknowledged. updated version is available (SAE J670e).

D13802 © IMechE 2004 Proc. Instn Mech. Engrs Vol. 218 Part D: J. Automobile Engineering

md00003802 11-02-04 14:46:57 Rev 14.05

The Charlesworth Group, Huddersfield 01484 517077

View publication stats